摘要本文主要研究基于纹理特征对焊接熔池图像进行分类，通过进行焊接工艺试验，对不同焊接参数进行改变来获得所需的熔池图像，主要是针对751类熔池图像进行基于纹理特征的图像分类，有最佳工艺，左偏，右偏，焊漏，气孔，保护气不足。单独采用灰度共生矩阵特征提取方法，基于改变参数值对图像灰度共生矩阵特征值熵，能量，逆矩阵，对比度，相关度，自相关数据统计整理，发现应用此算法分类结果比较理想，能较好的区分出在最佳工艺条件下与保护气不足时熔池图像，以及焊接左偏，右偏，焊漏，气孔的熔池图像，正确率可达85%。预先对熔池图像进行阈值分割轮廓提取，再进行GLCM特征提取，正确率提升到88%。对熔池图像进行纹理特征算法前预先提取轮廓，能更好的对熔池图像进行分类。33529
关键词 熔池图像； 纹理特征；阈值分割；灰度共生矩阵
毕业论文设计说明书外文摘要
Title Based on the texture characteristics of MAG welding molten pool image classification
Abstract
This paper mainly studies based on texture feature of the welding pool image classification, through welding experiment, the different welding parameters to
meet the need of change to get the molten pool images, mainly for six kinds of
molten pool images based on texture feature of image classification, optimum process, left, right, welding leakage, porosity, inadequate protection of gas.
Separately using gray level co-occurrence matrix feature extraction method, based
on changing the parameter value of image gray level co-occurrence matrix eight
value, entropy, energy, inverse matrix, contrast, correlation, since the relevant
Statistical data, found that applying this algorithm classification result is more
Ideal, can better distinguish between gas under the optimum technological conditions and protection when molten pool images, and welding to left, right,
Welding leakage, the porosity of the molten pool images, the accuracy can reach
85%. In advance of molten pool image threshold segmentation, contour extraction, then GLCM feature extraction, accuracy up to 88%. In advance before the molten
Pool image texture feature algorithm to extract the contour, can better classify
Molten pool images.
Keywords Weld pool image; Texture feature;Threshold segmentation; Gray symbiotic matrix.
目 次
1 绪论 1
1.1 选题背景及意义 1
1.2 焊接过程熔池视觉传感技术研究现状 2
1.3 熔池图像处理技术及特征提取研究现状 3
1.4 基于视觉的焊接质量智能控制研究现状 3
1.5 纹理特征提取研究现状 4
1.6 本课题研究内容 6
2 MAG焊熔池图像采集试验与图像处理分析 7
2.1 MAG焊熔池图像采集试验 7
2.1.1 试验设备与材料 7
2.1.2 MAG焊焊前准备与工艺试验设计方案制定 8
2.2 MAG焊熔池图像处理分析 10
3 MAG焊熔池图像基于纹理特征进行分类研究 13
3.1 基于灰度共生矩阵的熔池图像处理 13
3.1.1 基于灰度共生矩阵的熔池图像处理算法原理 13
3.1.2 熔池图像处理结果分析 13
3.2 基于阈值分割的熔池图像处理 14 基于纹理特征的MAG焊焊接缺陷熔池图像分类研究:http://www.751com.cn/cailiao/lunwen_30681.html